Fully digital spectrum analyzer using time compression and discrete fourier transform techniques

Claim 2

A spectrum analyzer comprising: converter means for receiving an analog signal and for providing successive sequences of N digital words f(n), each word being the n-th sample of a sequence of N successive samples of the analog signal (where n 1,2, . . .,N); first memory means for storing a sequence of N words provided by the converter means; window function multiplier means for combining each of the words from the first memory means with a selected window function W(n) to provide a sequence of N words f(n).W(n); second memory means for storing the sequence of N words provided by the window function multiplier means; trigonometric function multiplier means for combining the words from the second memory means with selected trigonometric functions to provide a plurality of series of N words R and N words I, where each R word results from combining a word f(n).W(n) with a corresponding cosine function of an argument including a value representing a spectral line K, and each I word results from combining a word f(n).W(n) with a corresponding sine function of an argument including a value representing said spectral line K, said R and I words being respectively the real and imaginary parts of Discrete Fourier Transforms of the sequence f(n).W(n), each of said plurality of series of R and I words being associated with a different one of a selected plurality of spectral lines K; means for accumulating, for each of the spectral lines K, the R words and the I words, and means for providing an average value of the N words R and an average value of the N words I; and means for generating, for each of said spectral lines K, the squares of said average values and for combining said squares to obtain a value representative of the power value for said spectral line K.

Claim 3

A spectrum analyzer as in claim 1 including means for averaging a plurality of successive power values for each spectral line K, said power values resulting from successive sequences of N words f(n), to thereby obtain a pluRality of averaged power values, one for each spectral line K.

Claim 4

A spectrum analyzer as in claim 2 including means for adding statistically controlled noise to the analog signal received by the converter means to enhance the resolution of the converter means.

Claim 5

A spectrum analyzer as in claim 3 including a digital source of sequences of words f(n) and means for selectively disconnecting the converter means from the first memory means and for storing in the first memory means a sequence of digital words f(n) from said digital source.

Claim 6

A spectrum analyzer as in claim 4 wherein the window function multiplier means includes means for storing a plurality of different window functions and for selectively combining each of the words from the first memory with a selected one of said stored window functions.

Claim 7

A spectrum analyzer as in claim 5 wherein the trigonometric function multiplier means includes trigonometric function storage means for storing trigonometric function values for one quadrant of a selected trigonometric function and control means for reading selected values from the trigonometric function storage means and for combining the last recited words with the f(n).W(n) words to thereby provide said R and I words.

Claim 8

A spectrum analyzer as in claim 6 wherein the means for averaging a plurality of power values includes means for averaging the last recited values in a sum mode.

Claim 9

A spectrum analyzer as in claim 7 wherein the means for averaging a plurality of power values includes means for averaging the last recited values in an exponential mode following said sum mode averaging thereof.

Claim 10

A spectrum analyzer as in claim 1 including a digital source of sequences of words f(n) and means for selectively disconnecting the converter means from the first memory means and for storing in the first memory a sequence of digital words f(n) from said digital source.

Claim 11

A spectrum analyzer as in claim 1 wherein the window function multiplier means includes means for storing a plurality of different window functions and for selectively combining each of the words f(n) from the first memory with a selected one of said stored window functions.

Claim 12

A spectrum analyzer as in claim 1 wherein the trigonometric function multiplier means includes means for storing trigonometric function values for a single quadrant of a selected trigonometric function and control means for reading selected values from the trigonometric function storage means and for combining the last recited values with the words f(n).W(n) to provide said R and I words.

Claim 13

A spectrum analyzer as in claim 1 including means for averaging a plurality of successive power values for each spectral line K, said power values resulting from successive sequences of N words f(n) to obtain a plurality of averaged power spectra values, one for each spectral line K, said averaging carried out in a sum mode.

Claim 14

A spectrum analyzer as in claim 12 wherein said averaging means includes means for additionally carrying out exponential mode averaging of said power values.

Claim 15

A spectrum analyzer as in claim 2 including means for adding statistically controlled noise to the analog signal received by the converter means.

Claim 16

A spectrum analyzer as in claim 14 wherein the noise comprises a noise signal which is substantially independent of the instantaneous value of the analog signal and has a selected amplitude probability and a selected power spectrum.

Claim 17

A spectrum analyzer having means for generating sequences of digital words, each sequence representing the time history of an analog signal over a selected time interval, means for generating from each sequence of digital words a set of signals representing the discrete Fourier transforms thereof for a selected set of spectral lines, and means for processing the last recited signals to provide signals indicative of the power spectrum of the analog signal, wherein the meAns for generating said signals representing discrete Fourier transforms comprises: means for combining the digital words with selected trigonometric functions to provide signals R and I for each digital word and each spectral line, where each signal R results from combining a digital word with a corresponding cosine function of an argument including a value representing a selected spectral line and each signal I results from combining a digital word with a corresponding sine function of an argument including a value representing the same selected spectral line, the resulting signals R and I being respectively the real and imaginary parts of the discrete Fourier transforms of the sequence of digital words for the selected set of spectral lines, wherein the combining means includes: means for storing a selected number of values for the same trigonometric function of an argument including a value representing each of said spectral lines; means for providing, for each different combination of a digital word and a spectral line, two generally different values from said storage means; and means for processing at least one of the provided values to transform it to a different trigonometric function of its arguments for providing thereby said sine function and said cosine function of the argument.

Claim 18

A spectrum analyzer having means for generating sequences of digital words, each sequence representing the time history of an analog signal over a selected time interval, means for generating from each sequence of digital words a set of signals representing the discrete Fourier transforms thereof for a selected set of spectral lines, and means for processing the last recited signals to provide signals indicative of the power spectrum of the analog signal, wherein the meAns for generating said signals representing discrete Fourier transforms comprises: means for combining the digital words with selected trigonometric functions to provide signals R and I for each digital word and each spectral line, where each signal R results from combining a digital word with a corresponding cosine function of an argument including a value representing a selected spectral line and each signal I results from combining a digital word with a corresponding sine function of an argument including a value representing the same selected spectral line, the resulting signals R and I being respectively the real and imaginary parts of the discrete Fourier transforms of the sequence of digital words for the selected set of spectral lines, wherein the combining means includes: means for storing a selected number of values for the same trigonometric function of an argument including a value representing each of said spectral lines; means for providing, for each different combination of a digital word and a spectral line, two generally different values from said storage means; and means for processing at least one of the provided values to transform it to a different trigonometric function of its arguments for providing thereby said sine function and said cosine function of the argument.

Claim 19

A spectrum analyzer as in claim 16 wherein the storing means stores only the values for said trigonometric function for one quadrant of a wave cycle wave.

Claim 20

A spectrum analyzer having means for generating sequences of digital words, each sequence representing the time history of an analog signal over a selected time interval, means for generating from each sequence of digital words a set of signals representing the discrete Fourier transforms thereof for a selected set of spectral lines, and means for processing the last recited signals to provide signals indicative of the power spectrum of the analog signal, wherein the means for generating said signals representing discrete Fourier transforms comprises: means for combining the digital words with selected trigonometric functions to provide signals R and I for each digital word and each spectral line, where each signal R results from combining a digital word with a corresponding cosine function of an argument including a value representing a selected spectral line and each signal I results from combining a digital word with a corresponding sine function of an argument including a value representing the same selected spectral line, the resulting signals R and I being respectively the real and imaginary parts of the discrete Fourier transforms of the sequence of digital words for the selected set of spectral lines, wherein each of the digital words of a sequence has a rank and each spectral line has a rank, and wherein the combining means includes: means for storing a selected number of values of a single trigonometric function of an argument including a value representing each of the selected spectral lines; means for combining a representation of the rank of each digital word with a representation of the rank of each spectral line to provide addresses identifying values in the storing means and to provide an indication of the type of trigonometric function required for combining with the digital words and of the quadrant thereof; means for providing from the storing means the values identified by the last recited addresses; and means for processing the values provided from the storing means in accordance with said indications to provide the recited sine functions and cosine functions.

Claim 21

A spectrum analyzer having means for generating sequences of digital words, each sequence representing the time history of an analog signal over a selected time interval, means for generating from each sequence of digital words a set of signals representing the discrete Fourier transforms thereof for a selected set of spectral lines, and means for processing the last recited signals to provide signals indicative of the power spectrum of the analog signal, wherein the means for generating said signals representing discrete Fourier transforms comprises: means for combining the digital words with selected trigonometric functions to provide signals R and I for each digital word and each spectral line, where each signal R results from combining a digital word with a corresponding cosine function of an argument including a value representing a selected spectral line and each signal I results from combining a digital word with a corresponding sine function of an argument including a value representing the same selected spectral line, the resulting signals R and I being respectively the real and imaginary parts of the discrete Fourier transforms of the sequence of digital words for the selected set of spectral lines, wherein each of the digital words of a sequence has a rank and each spectral line has a rank, and wherein the combining means includes: means for storing a selected number of values of a single trigonometric function of an argument including a value representing each of the selected spectral lines; means for combining a representation of the rank of each digital word with a representation of the rank of each spectral line to provide addresses identifying values in the storing means and to provide an indication of the type of trigonometric function required for combining with the digital words and of the quadrant thereof; means for providing from the storing means the values identified by the last recited addresses; and means for processing the values provided from the storing means in accordance with said indications to provide the recited sine functions and cosine functions.

Claim 22

A spectrum analysis method comprising the steps of: receiving an analog signal and providing successive sequences of N digital words f(n), each word being the n-th sample of a sequence of N successive samples of the analog signal (where n 1, 2, . . . , N); storing a sequence of N of said digital words f(N); combining each of the stored words f(n) with a selected window function W(n) to provide a sequence of N words f(n).W(n); storing the sequence of N words f(n).W(n); combining the last recited words with selected trigonometric functions to provide a plurality of series of N words R and N words I, where each R word results from combining a word f(n).W(n) with a corresponding cosine function of an argument including a value representing a spectral line K, and each I word results from combining a word f(n).W(n) with a corresponding sine function of an argument including a value representing said spectral line K, said R and I words being respectively the real and imaginary parts of Discrete Fourier Transforms of the sequence f(n).W(n), each of said plurality of series of R and I words being associated with a different one of a selected plurality of spectral lines K; accumulating, for each of the spectral lines K, the R words and the I words, and providing an average value of the N words R and an average value of the N words I; and generating, for each of said spectral lines K, the squares of said average values and combining said squares to obtain a value representative of the power value for said spectral line K.

Claim 23

A spectrum analysis method comprising the steps of: receiving an analog signal and providing successive sequences of N digital words f(n), each word being the n-th sample of a sequence of N successive samples of the analog signal (where n 1, 2, . . . , N); storing a sequence of N of said digital words f(N); combining each of the stored words f(n) with a selected window function W(n) to provide a sequence of N words f(n).W(n); storing the sequence of N words f(n).W(n); combining the last recited words with selected trigonometric functions to provide a plurality of series of N words R and N words I, where each R word results from combining a word f(n).W(n) with a corresponding cosine function of an argument including a value representing a spectral line K, and each I word results from combining a word f(n).W(n) with a corresponding sine function of an argument including a value representing said spectral line K, said R and I words being respectively the real and imaginary parts of Discrete Fourier Transforms of the sequence f(n).W(n), each of said plurality of series of R and I words being associated with a different one of a selected plurality of spectral lines K; accumulating, for each of the spectral lines K, the R words and the I words, and providing an average value of the N words R and an average value of the N words I; and generating, for each of said spectral lines K, the squares of said average values and combining said squares to obtain a value representative of the power value for said spectral line K.

Claim 24

A method as in claim 19 including the further step of averaging a plurality of successive power values for each spectral line K, said power values resulting from successive sequences of N words f(n), to thereby obtain a plurality of averaged power spectra values, one for each spectral line K.

Claim 25

A method as in claim 19 wherein the step of combining each word f(n) with a window function W(n) comprises storing a plurality of different window functions and selectively combining each of the words f(n) with a selected one of said stored window functions.

Claim 26

A method as in claim 19 wherein the step of combining the words f(n).W(n) with selected trigonometric functions comprises storing trigonometric function values for one quadrant of a selected trigonometric function and combining the last recited values with the f(n).W(n) words to thereby provide said R and I words.

Claim 27

A method as in claim 19 including the further step of averaging a plurality of the obtained power values in a sum mode.

Claim 28

A spectrum analyzer as in claim 23 including the further step of averaging a plurality of the obtained power values in an exponential mode following said sum mode averaging.

Claim 31

A method as in claim 25 wherein the storing step includes storing only the value for said trigonometric function for one quadrant of a wave cycle.